#include <ocl.hpp>
Public Member Functions | |
KalmanFilter () | |
KalmanFilter (int dynamParams, int measureParams, int controlParams=0, int type=CV_32F) | |
the full constructor taking the dimensionality of the state, of the measurement and of the control vector More... | |
void | init (int dynamParams, int measureParams, int controlParams=0, int type=CV_32F) |
re-initializes Kalman filter. The previous content is destroyed. More... | |
const oclMat & | predict (const oclMat &control=oclMat()) |
const oclMat & | correct (const oclMat &measurement) |
Public Attributes | |
oclMat | statePre |
predicted state (x'(k)): x(k)=A*x(k-1)+B*u(k) More... | |
oclMat | statePost |
corrected state (x(k)): x(k)=x'(k)+K(k)*(z(k)-H*x'(k)) More... | |
oclMat | transitionMatrix |
state transition matrix (A) More... | |
oclMat | controlMatrix |
control matrix (B) (not used if there is no control) More... | |
oclMat | measurementMatrix |
measurement matrix (H) More... | |
oclMat | processNoiseCov |
process noise covariance matrix (Q) More... | |
oclMat | measurementNoiseCov |
measurement noise covariance matrix (R) More... | |
oclMat | errorCovPre |
priori error estimate covariance matrix (P'(k)): P'(k)=A*P(k-1)*At + Q)*/ More... | |
oclMat | gain |
Kalman gain matrix (K(k)): K(k)=P'(k)*Ht*inv(H*P'(k)*Ht+R) More... | |
oclMat | errorCovPost |
posteriori error estimate covariance matrix (P(k)): P(k)=(I-K(k)*H)*P'(k) More... | |
***************Kalman Filter*************!
cv::ocl::KalmanFilter::KalmanFilter | ( | ) |
cv::ocl::KalmanFilter::KalmanFilter | ( | int | dynamParams, |
int | measureParams, | ||
int | controlParams = 0 , |
||
int | type = CV_32F |
||
) |
the full constructor taking the dimensionality of the state, of the measurement and of the control vector
void cv::ocl::KalmanFilter::init | ( | int | dynamParams, |
int | measureParams, | ||
int | controlParams = 0 , |
||
int | type = CV_32F |
||
) |
re-initializes Kalman filter. The previous content is destroyed.
oclMat cv::ocl::KalmanFilter::controlMatrix |
control matrix (B) (not used if there is no control)
oclMat cv::ocl::KalmanFilter::errorCovPost |
posteriori error estimate covariance matrix (P(k)): P(k)=(I-K(k)*H)*P'(k)
oclMat cv::ocl::KalmanFilter::errorCovPre |
priori error estimate covariance matrix (P'(k)): P'(k)=A*P(k-1)*At + Q)*/
oclMat cv::ocl::KalmanFilter::gain |
Kalman gain matrix (K(k)): K(k)=P'(k)*Ht*inv(H*P'(k)*Ht+R)
oclMat cv::ocl::KalmanFilter::measurementMatrix |
measurement matrix (H)
oclMat cv::ocl::KalmanFilter::measurementNoiseCov |
measurement noise covariance matrix (R)
oclMat cv::ocl::KalmanFilter::processNoiseCov |
process noise covariance matrix (Q)
oclMat cv::ocl::KalmanFilter::statePost |
corrected state (x(k)): x(k)=x'(k)+K(k)*(z(k)-H*x'(k))
oclMat cv::ocl::KalmanFilter::statePre |
predicted state (x'(k)): x(k)=A*x(k-1)+B*u(k)
oclMat cv::ocl::KalmanFilter::transitionMatrix |
state transition matrix (A)